Post on 05-Oct-2020
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Math 3313: Differential EquationsFirst-order ordinary differential equations
Thomas W. Carr
Department of MathematicsSouthern Methodist University
Dallas, TX
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Outline
Math 2343: Introduction
Separable ODEs
Linear, non-homogeneous
Graphical analysis
Numerical Approximation
Applications
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Outline
Math 2343: Introduction
Separable ODEs
Linear, non-homogeneous
Graphical analysis
Numerical Approximation
Applications
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
What’s it about?
The change in x(t) w/rt t is given by the function f (x , t).
dx(t)dt
= f (x , t) (1)
Goals of this class.• What is x(t)?
Solution methods?• Where did the differential equation come from?
Modeling• Interpretation
What does the solution x(t) say about the “physics”?
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Some terms
Ordinary Differential Equation (ODE)Has only 1 independent variable t .
dxdt
= f (x , t) (2)
Partial Differential Equation (PDE)Has only 2 or more independent variables (ex. time & space).
∂x∂t
=∂2x∂z2 (Diffusion) (3)
1st order ODE→ highest derivative is 1. dxdt = f (x , t)
2nd order ODE→ highest derivative is 2. d2xdt2 = f (x , dx
dt , t)
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Linear vs. Nonlinear
Linear nth order
an(t)dnxdtn + an−1(t)
dn−1xdtn−1 + . . .+ a1(t)
dxdt
+ a0(t)x = f (t) (4)
x , dxdt , . . . dnx
dtn appear linearly.No x2, sin(x), x dx
dt , . . .
t doesn’t matter. t2, sin(t), . . . are OK.
Nonlineardxdt
= sin(x),(
dxdt
)2
+ xdxdt
= 2.
sin(x), ( dxdt )
2, x dxdt are nonlinear functions of x .
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Solutions must satisfy the ODE
ex. Is x(t) = ce−3t a solution to
dxdt
= −3x ? (5)
Substitute and check!
ex. Is x(t) = c1 sin2t + c2 cos2t a solution to
d2xdt2 + 4x = 0? (6)
Substitute and check!
Make it a habit to check!Find x then substitute.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Explicit vs. Implicit
Explicit: when we can solve for x = F (t).
Implicit: Left with G(x(t), t) = 0.
You don’t have x = something. Instead, it’s defined implicitly by thefunction G.
ex. Let G(x , t) = x3 + x − t + 1t + c. Show that G = 0 is a solution to
t2 dxdt
=t2 + 1
3x2 + 1(7)
Substitute and check!
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Solving implies "integrating"ex.
dxdt
= sin t (1st order, linear) (8)
What function x has derivative sin t?
x(t) =∫
sin t dt + c = − cos t + c.
ex.
d3xdt3 = 1 (3rd order, linear) (9)
d2xdt2 = t + c1
dxdt
=12
t2 + c1t + c2
x =16
t3 +c1
2t2 + c2t + c3.
Every time we integrate we pick up a constant.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
The constants and Initial Conditions
The solution to an nth order ODE will have n constants.Specify the cj with n initial conditions (ICs)
ex.dxdt
= cos γt AND x(t =π
2γ) = 3. (10)
Integrate and solve for c!
The general solution: x(t) = 1γ sin γt + c
The solution to the initial value problem: x(t) = 1γ sin γt + (3− 1
γ )
ex.dxdt
= tet2, x(0) = 1. (11)
Integrate (definite vs. indefinite integration)
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Outline
Math 2343: Introduction
Separable ODEs
Linear, non-homogeneous
Graphical analysis
Numerical Approximation
Applications
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
What is "separable"?Given
dxdt
= f (t) Integrate
dxdt
= f (x , t) Need other methods (if at all doable)
dxdt
= f (x , t) = g(x)h(t) SEPARABLE
f (x , t) is the product of a function of x (g) with a function of t (h).
ex.
f (x , t) = tx → g(x) = x , h(t) = tf (x , t) = x2et → g(x) = x2, h(t) = et
f (x , t) = sin(xt) → ???
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Separation of variables
1. Given:
dxdt
= g(x)h(t)
2. Separate x and t :
1g(x)
dxdt
= h(t)
3. Relabel 1/g = p(x):
p(x)dxdt
= h(t)
4. Integrate w/rt t∫p(x)
dxdt
dt =∫
h(t)dt
5. Integrate w/ substitution
Let u = x(t)⇒dudt
=dxdt⇒ du =
dxdt
dt
∫p(u)du =
∫h(t)dt
antider of p(u)|u=x = antider of h(t)
May or may not be able to integrate.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Separation: examples
ex. Find the general solution to
dxdt
= −6tx (12)
Separate and solve.
ex.
dxdt
= (x − 1)2 sin t (13)
Separate and solve.
ex. Find the solution to the initial value prob-lems below.
t2 dxdt
=t2 + 1
3x2 + 1, x(1) = 2. (14)
Separate and solve.
ex.dxdt
= tet2+x2, x(t0) = x0 (15)
Separate and solve.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Linear, constant coefficient
Special case of separable ODE
dxdt
= kx (16)
Linear: dxdt and x are linear functions of x .
Constant coefficient: 1 and k .
1x
dx = kdt
ln x = kt + cx = ekt+c
x = ecekt
x = c̃ekt
Linear, constant-coefficient ODEs always have ert as solutions.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Linear, constant coefficient: examples
ex.dxdt
= 5x (17)
Substitute and find r .
ex.dxdt− kx = 0 (18)
Substitute and find r .
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Outline
Math 2343: Introduction
Separable ODEs
Linear, non-homogeneous
Graphical analysis
Numerical Approximation
Applications
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
What is "linear, nonhomogenous"?Given
dxdt
= f (t) Integrate
dxdt
= f (x , t) = g(x)h(t) separate
dxdt
= f (x , t) = kx Const. coeff. → ert
Nowdxdt
= f (x , t) ⇒ dxdt
+ p(t)x = f (t) (19)
p(t): variable coefficient. Function of t .f (t): nonhomogenous/forcing term. Function of t .
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Some properties
The f (t) prevents separation.
1x
dxdt
= −p(t) +f (t)x
The solution to a linear nth order nonhomog. ODE has 2 parts.
x(t) = xh(t) + xp(t) = homogeneous sol. + particular sol.
xh :dxh
dt+ p(t)xh = 0 (f = 0).
xp :dxp
dt+ p(t)xp = f (t)
Check. Let x = xh + xp. Substitute into original. Use the above tocancel terms.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Integrating factor method
Put ODE into standard form
Step 1:dxdt
+ p(t)x = f (t)
Multiply by I.F. u(t) (unknown)
u(t)dxdt
+ u(t)p(t)x = u(t)f (t)
Choose u such that
u(t)p(t) =dudt
Make replacement
u(t)dxdt
+dudt
x(t) = u(t)f (t)
This is the result of product rule.
ddt
(ux) = udxdt
+dudt
x
Make replacement
Step 3:ddt
(u x) = uf
Integrate
Step 4:∫
ddt
(u x) dt =∫
u f dt
u x =
∫u fds + c
Solve for x .
The method:Choose u so we can integrate.
dudt
= p(t)u.
1u
du = p(t)dt
Step 2: u = e∫
p(t)dt
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
A first example
ex.dxdt
= 3x + e2t , x(0) = 5
Apply IF method.
Summary of IF Method1. Standard form: dx
dt + p(t)x = f (t)
2. If is: u = e∫
p(t)dt3. ODE becomes:
∫ ddt (u x) dt =
∫u f dt
4. Integrate
Turn something you didn’t know how to solve into something you do (integration). Priceis needing to find u(t).
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Some more IF examples
ex.t2 dx
dt+ tx = t sin t , x(1) = 2 (20)
Apply IF method.
ex.dxdt
+ t4x = 1 (21)
Apply IF method.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Outline
Math 2343: Introduction
Separable ODEs
Linear, non-homogeneous
Graphical analysis
Numerical Approximation
Applications
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Online analysis tools
• GeoGebra Slope Field Plotter:www.geogebra.org/m/W7dAdgqc
• Blufton Univ. Slope and Direction Fields:bluffton.edu/homepages/facstaff/nesterd/java/slopefields.html
• Interactive Differential Equations:www.aw-bc.com/ide/
First, some board work.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Direction Fields
dxdt
= t2
• f (t , x) = t2 Non-negative.x(t) never decreases.
• f (t , x) = 0? t2 = 0?⇒ t = 0.When t = 0, x(t) is horizontal.
• f (t , x) = 1?: t2 = 1?⇒ t = ±1.When t = ±1, x(t) has slope 1.
• Integrate to find solution:x(t) = 1
3 t3 + c.
Note missing axis labels!(Blufton’s Slope and Dir. Fields tool)
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Direction Fields
dxdt
= x2 − 1
• If x < −1 or x > 1, thenf > 0 so x increases.
• If −1 < x < 1, thenf < 0 so x decreases.
• If x = ±1, thenf = 0 so x is at Equilibrium.
• Solve by Sep of Varx(t) = 1+ce2t
1−ce2t
• Singular when 1− ce2t = 0.Finite time blowup. Note missing axis labels!
(Blufton’s Slope and Dir. Fields tool)
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Autonomous ODEs & EquilibriumFirst, some board work.
0 1 2 3 4 5 6 7 8 9 10
−3
−2
−1
0
1
2
3
t
x
x ’ = x2 − 1
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Autonomous ODEs: example 2First, some board work.
0 1 2 3 4 5 6 7 8 9 10
−3
−2
−1
0
1
2
3
t
x
x ’ = x − x3
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Autonomous ODEs: example 3
0 1 2 3 4 5 6 7 8 9 10
−10
−8
−6
−4
−2
0
2
4
6
8
10
t
x
x ’ = sin(x)
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Existence & UniquenessBefore we start solving....
• How do we know if there is a solution to find? Existence.• If we find a solution, how do we know if it is the only one?
Uniqueness.ex.ODE: dx
dt = −x ⇒ x(t) = ce−t : (a family of solution curves)IC: x(t0) = x0 ⇒ c = x0et0 ⇒ x(t) = x0et0−t
(a specific curve passing through a specific point (t0, x0))
Every point (t , x) has one and only one solution curve passingthrough it.
• If no solution curve: Does not exist.• If more than one: Not unique.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
E& U: Theorem
Given dxdt = f (x , t). If both f (x , t) and ∂f
∂x (x , t) are continuous in aregion containing (t0, x0), then there exists a unique solution through(t0, x0). (f and its partial with respect to x must be continuous.)
ex.dxdt
=1t
(22)
f is discontinuous at t = 0 so theorem fails at t = 0.Solve: x(t) = ln |t |+ c. Undefined at t = 0.
For t 6= 0, where the theorem is satisfied, there is one and only onesolution through each point.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
E& U: Example
ex.dxdt
=xt
(23)
Examine f and ∂f/∂x . Then solve.
ex.dxdt
=√
x2 − t2 (24)
f (x , t) must be real. Simulate.
ex.dxdt
= x2/3 vs.dxdt
= x4/3 (25)
Simulate and compare.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Outline
Math 2343: Introduction
Separable ODEs
Linear, non-homogeneous
Graphical analysis
Numerical Approximation
Applications
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Analysis vs. Numerics/Simulation
Givendxdt
= f (x , t), x(t0) = x0
• What if we can’t separate, can’t use an I.F., isn’t Exact (see text)?• What if the problem is too complicated and an analytical solution
is not possible? Which is MOST!• Use computer simulations to find a numerical approximation.
But computers can’t do "calculus". How do we approximate theproblem so that computers can operate on it?
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Linear approximation
Given a curve x(t). Suppose you know a point on the curve (t0, x0) The linearapproximation (tangent line) is
xl (t)− x0 = m(t − t0), where m =dxdt
(t0). (26)
xl (t) = x0 +dxdt
(t0)(t − t0)
Use the line as an approximation to thecurve. At t = t1:
• True value is x(t1).
• Approximate value is xl (t1) = x1.
• Error: e1 = x(t1)− x1
If the step size h = t1 − t0 is not too big, we expect the error to not be too big.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Euler’s methodGiven:
dxdt
= f (x , t), x(t0) = x0
To get x(t1) use the approximation x1:
x1 = x0 +dxdt
(t0)(t1 − t0)
The derivative is given by the ODE:
x1 = x0 + f (x0, t0)(t1 − t0)
Go to a new point when t = t2.Use the linear approx. again.
x2 = x1 + f (x1, t1)(t2 − t1)
Repeat, repeat, . . .
xn+1 = xn + f (xn, tn)(tn+1 − tn)
If fixed stepsize: h = tn+1 − tn.
xn+1 = xn + h fn (27)
• f is given by the ODE, so known.
• You pick the times tn, so known.
• The xn are repeated approxs.
• Approx. based on approx!
While x0 = x(t0), x1 6= x(t1).
Error Euler’s method is "Order h"
|en| = |x(tn)− xn| = Mh
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Euler examples
For each problem below:• Solve analytically then evaluate at the specified points.• Solve numerically using Euler’s method.
ex.dxdt
= 5 + 2x , x(0) = 0, t = 0, 0.1, 0.2, 0.3
Compare analytical and numerical solutions.
ex.dxdt
= 3x2, x(0) = 1, t = 0, 0.2, 0.4, 0.6
Compare analytical and numerical solutions.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Issues
• To go long times, need more steps. Error can accumulate.
• To reduce error, reduce the stepsize h. Now computer takes a long time.
• Perhaps better methods. Instead of using linear (tangent) approx, use quadraticapprox, or polynomial approx, of weighted averages of derivatives, or . . . .
• Implement error "correction." Take a step, estimate error, devise a scheme toeliminate the error.
• Better methods and error correction require more work by the computer. Now thecomputer takes more time.
• Buy a faster computer.
• If you don’t have an analytical solution to compare against, how do you know thenumerical method gives a correct result? What is the numerical solutionconverging to?
• Issues, caveats, issues, caveats, issues,. . .
• A deep knowledge of these issues and solutions, i.e., the fields of ScientificComputer and Numerical Analysis, gets you jobs.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
euler.m
dxdt = 5 + 2x, x(0) = 0
Default: h = 0.1, N = 100, tf = 10.Change h = 0.2 ⇒ tf = 20.
Exp. growing solution gets large.Change h = 0.1, N = 10 for better view.Change h = 0.2, N = 5.
Can see tangent lines.Change h = 0.3, N = 4.Note:
• When h = 0.1, x(1) ≈ 13.
• When h = 0.3, x(1) ≈ 10.
• Which is more accurate?x(1) = − 5
2 + 52 e2 = 15.9 . . ..
Change h = 0.01, N = 100.
Very accurate but more steps. (slow?)
dxdt = 5 − 2x, x(0) = 0
What do we expect? Always ask yourself,what do you expect?
Default: h = 0.1, N = 100, tf = 10.Goes to steady state at 5
2 .Change h = 0.2.Change h = 0.5. Not smooth.Change h = 0.6. Overshoot.Change h = 0.8. Oscillations.Change h = 1.0. UNSTABLE!
• Accuracy issues.
• Stability issues.
• We can do better thanEuler.
• Different methodshave differentadvantages anddisadvantages.
• In matlab, ode23 andode45 are goodall-purpose solvers.
• Interested in details?MATH 3315
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Outline
Math 2343: Introduction
Separable ODEs
Linear, non-homogeneous
Graphical analysis
Numerical Approximation
Applications
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Modeling process
Given some problem to solve.• May be from science,
engineering, economics,finance,. . .
Model (describe) the problem with a dif-ferential equation.
• This can be quite difficult andtime consuming.
Solve the DE-model• Exactly
• Approximately
• Numerically/computationally
• All of the above
Evaluate the solution• Does the solution described
previously observed behavior?
• Should the model be modified?
• Is the model good in somerestricted set of cases?
• Can the model predict behaviornot yet observed?
Are there other ways to model theproblem? DEs are just one tool ofmany.
Philosophy on the application of math-ematics:
• It is not “exact.”
• Requires judgment andimagination.
• Requires knowledge of both theapplication and mathematics.
• Requires collaboration andcommunication acrossdisciplines.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Growth/Decay Rate
There is a branch of mathematical biology called “PopulationDynamics,” where the competition between species is studied. This isimportant in environmental resource management. “Epidemiology” isvery similar in that there is a competition between those who aresusceptible, those who are infected and those who are recovered,from a particular disease.
dPdt
= Rate of change of the population.
dPdt
P=
1P
dPdt
=Rate of change
population= Growth/Decay Rate
Growth rate measures the Rate of change with respect to thepopulation size, i.e., the relative rate of change. The distinction isimportant.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Exponential vs. linearSuppose dP
dt = k .• There are always k new
individuals in a given time,independent of P. For example,no matter how many have beenadmitted, Bush Stadium’sturnstile gates allow only a fixednumber to enter over a giventime.
• Solve by integration.
• P(t) = kt + C, apply ICC = P(0), so P(t) = P(0) + kt .
• Results in linear growth.
Suppose 1P
dPdt = k or dP
dt = kP.• The rate of change depends on
the population size. For example,this reflects the fact that withmore members in a populationthere can be more births.
• Solve by separation.
• P(t) = Cekt , apply IC C = P(0),so P(t) = P(0)ekt .
• Results in exponential grown
Which model is more accurate dependson the application.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Decay and Death
Decay is just the negative of growth.dPdt = −k so P(t) = P(0)− kt .• Linear decrease.• People leave Bush Stadium at a fixed rate.
dPdt = −kP so P(t) = P(0)e−kt .• Exponential decay.• The more members of a population, the more deaths there are.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Money in savings account
P = amount (principal). dPdt = change in time.
How can the amount change?
• Interest: "Interest rate" = "Growth rate" = r1
dPdt
= "interest rate" times P
= r1P⇒ P(t) = P0ert
• Deposits & withdrawals: on average, r2 dollars/day.
dPdt
= r2
⇒ P(t) = r2t + P0
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Money in savings account
Together.... Interest AND deposits+withdrawals?
• Must start with new ODE.• DO NOT ADD THE SOLUTIONS FROM ABOVE.
New ODE with both proceses.
dPdt
= r1P + r2, P(0) = P0 (28)
Solve and note effect of compound interest!
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
A few more applications
Radioactive decayExperimental observation: The rate of decay of a radioactive materialis proportional to the number of atoms present.
• Half-life• Doubling time (money in the bank)
Model and solve
Single-species populationBirth, death, deposits and withdrawals.
• Equilibrium• Fish are positive
Model and solve
Argon LASER
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
And two more applications
Your money-market checking account comes with an interest rate of2% (r = 0.02 1/day). On average you withdray $3 dollars/day. Initially,you have $1000 in the account. When does your balance increasetenfold?Model and solve.
Thermal cooling: the rate of change of the surface temperature of anobject is proportional to the difference between the temperature of theobject and it’s surroundings. (Newton’s law of cooling)Modeling and solving.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
And mixture (tank) applicationsConcerned with the amount of a given substance present in a solutionas a function of time. Our goal is to formulate and solve a differentialequation for the quantity Q(t) of interest. Consider a box of sand:
dQ
dt
dQ
dt
IN
Q = amount of sand
= ?
IN 1
IN 2
IN 3
OUT 1
OUT 2
OUT
= INs − OUTs
LAW of MASS BALANCEThe rate of change of Q = rate of sand in - rate of sand out.
dQdt
=
[dQdt
]in−[
dQdt
]out
Similar to dPdt = (Births - Deaths) or (Deposits - Withdrawals).
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
A problem with just flow
First, a simple example with just water, not a mixture.ex. Consider a tank that can hold 1000 gal of water. Water is beingpumped into the tank at a rate of 10 gal/min. Water is pumped out ofthe tank at rate of 8 gal/min. Initially, there are 200 gal in the tank.Formulate and solve an ODE for the amount of water in the tank.
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Now with stuff in the flow
Suppose there is "stuff" mixed into the water, i.e., there is aconcentration of "stuff" in the volume of water. How do we determinethe amount quantity of "stuff" that is in the mixture?
Concentration =QuantityVolume
⇒ C(t) =Q(t)V (t)
Both the amount Q(t) and the volume V (t) may be functions of time.Hence, the concentration C(t) also changes in time.
How do we get the rate of change of the quantity/amount of stuff?
dQdt
= flow rate · concentration (29)
masstime
=volume
time· mass
volume
Math 2343: Introduction Separable ODEs Linear, non-homogeneous Graphical analysis Numerical Approximation Applications
Two mixture problems
ex. A room has a volume of 800 ft3. The air in the room containschlorine at an initial concentration of 0.1 g/ft3. Fresh air enters theroom at a rate of 8 ft3/min. The air in the room is well mixed and flowsout of the door at the same rate that it flows in.Find the concentration of chlorine as a function of time.
ex. A well circulated pond contains 1 million L of water. It containspollutant at a concentration of 0.01 kg/L. Pure water enters from astream at 100 L/h. Water evaporates from the pond (leaving thepollutant behind) at 10 L/h and water flows out a pipe at 80 L/h.How many days will it take for the pollution concentration to drop to0.001 kg/L?